Professional Documents
Culture Documents
Finite Time SDRE Control of F16 Air
Finite Time SDRE Control of F16 Air
142848
Archives of Control Sciences
Volume 32(LXVIII), 2022
No. 3, pages 557–576
This paper proposes a finite-time horizon suboptimal control strategy based on state-
dependent Riccati equation (SDRE) to control of F16 multirole aircraft. Flight stabilizer control
of super maneuverable aircraft is modelled and simulated. For aircraft modelling purpose a
full 6 DOF model is considered and described by nonlinear state-space approach. Also a stable
state-dependent parametrization (SDP) necessary for solution of the SDRE control problem is
proposed. Solution of the SDRE control problem with adequate defined weighting matrices
in performance index shows possibility of fast and optimal aircraft control in finite-time. The
method in this form can be used for stabilization of aircraft flight and aerodynamics.
Key words: aircraft modelling, state-dependent Riccati equations, finite-time optimal con-
trol
1. Introduction
that is the design challenge and has caused most problems [1, 29, 36]. The re-
lationship between pilot action and aircraft reaction should be fast, performed
in finite time, eliminating aerodynamics nonlinearities, uncertainties or uncon-
trollable oscillations resulting from efforts of the pilot to control the aircraft and
occur when the pilot of an aircraft inadvertently commands an often increas-
ing series of corrections in opposite directions, known as PIO (Pilot Involved
Oscillations) [19, 36].
One of popular and very effective is a single-engine multirole fighter aircraft
F16 originally developed by General Dynamics for the United States Air Force
(USAF) also used by Polish Air Force (PAF). The multirole F16 aircraft is shown
in Fig. 1.
Many papers deal with the F16 aircraft dynamics, where finite-time opti-
mal control methods used to flight control systems is still a challenge for many
engineers and researchers [1,18,22,23,29,34,35]. Literature provide some inter-
esting works related to finite-time problems. In [12] authors studied a finite-time
sliding mode attitude controller for a reentry vehicle with blended aerodynamic
surfaces and a reaction control system. Next work [29] investigated the path
planning of a reusable launch vehicle using a finite-horizon suboptimal con-
troller. Further, the paper [9] deals with digital controller for longitudinal aircraft
model, where the control task is formulated as a tracking problem of velocity and
flight path angle considering incomplete information about varying parameters
of the system and external disturbances. The output tracking control, employ-
ing model reference feedback linearization of an aircraft subject to additive,
uncertain, nonlinear disturbances is presented and described in [27]. An inter-
esting control problem is studied in [31] where the finite-time attitude tracking
FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 559
control problem of a reusable launch vehicle in the reentry phase under in-
put constraints using a constrained adaptive back-stepping fast terminal sliding
mode control technique. A nonsingular terminal sliding-mode control method
with finite-time fault-tolerant control for spacecraft with actuator saturations is
studied in [18].
Nowadays, modern optimal control theory proposes high performance and
rapidly emerging control technique called finite-time SDRE (State-Dependent
Riccati Equation) [2, 7, 20, 24, 32]. This is a suboptimal control methodology for
nonlinear systems. The technique uses direct parameterization to bring the nonlin-
ear system to a linear structure having state-dependent coefficients (SDC) [33].
A state-dependent Riccati equation (SDRE) is then solved accordingly to the
change of state trajectory to obtain a nonlinear feedback controller matrix, which
coefficients, in other feedback gains are the solution of SDRE [8].
The method, firstly proposed in 1962 [21] and later expanded in 1975 [32],
was further carefully analyzed and deeply studied creating its useful form for
technical applications [20]. The method employs parameterization of the nonlin-
ear dynamics into the state vector, then the product of a matrix-valued function
depends on the state itself [33]. The control technique fully captures the non-
linearities of the dynamic system, bringing the system to a (nonunique) linear
structure having state-dependent coefficient (SDC) matrix form, and minimiz-
ing also a state-dependent nonlinear performance index having a quadratic-like
structure. The differential SDRE equations using the SDC matrices is then solved
on-line (in real time) to give the suboptimum control law. The technique for
the finite-time nonlinear optimal control problem in the multivariable case is
locally asymptotically stable and locally asymptotically optimal as described in
following theoretical contributions [2, 4, 6, 16, 20].
Applications of the SDRE control technique include also satellite and space-
craft control and estimation, integrated guidance and control design, autopilot
design, robotics, control of systems with parasitic effects, control of artificial
human pancreas, ducted fan control and magnetic systems including levitation
and drives [2, 7, 18].
In this paper, a modelling and control design methodology as the concept
is proposed to design of high-performance and optimal flight stabilizer for F16
multitask military aircraft. The paper presents a nonlinear model of the aircraft
and solution of the finite-time suboptimal control problem for flight stabilization
problem minimizing energy lost and delivered to the flaying machine.
The rigid body equations of motion are the differential equations that describe
the evolution of basic states of an aircraft. The aircraft model presents Fig. 2.
560 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ
The aircraft dynamics is generally defined using Newton’s force and moment
equations [34, 35]. The force equation is following
F = 𝑚 (¤v + 𝝎 × v) , (1)
¤ + 𝝎 × I𝝎,
M = I𝝎 (2)
where I is an aircraft inertia matrix and M denotes moment vector. When consider
vector v defined for all components in 𝑥, 𝑦 and 𝑧 direction and 𝝎 for roll 𝜑, pitch
𝜃 and yaw 𝜓 angle
" # " #
𝑢 𝑝
v= 𝑣 and 𝝎 = 𝑞 (3)
𝑤 𝑟
then equations of aircraft aerodynamics can be defined for linear and angular
speeds. In addition, because of a plane of symmetry so in the inertia matrix the
cross-products involving 𝑦 become zero
𝐼𝑥𝑥 0 −𝐼𝑥𝑧
" #
I = 0 𝐼 𝑦𝑦 0 . (4)
−𝐼 𝑧𝑥 0 𝐼 𝑧𝑧
FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 561
The system of nonlinear equations that describes aircraft flight dynamics, con-
sidering gravity forces 𝑔 and force due to the thrust 𝐹𝑇 , is following
is the nonlinear optimal gain. In order to extract the SDRE-like form of equation
(19), it is assumed that K(x) is the suboptimal solution to
¤
K(x) + K(x)A(x) + A(x)𝑇 K(x)
− K(x)B(x)R(x) −1 B(x)𝑇 K(x) + Q(x) = 0 (20)
optimal gain. However, the numerical difficulties when both solutions are used
become problematic in this technique, due to computation of co-state vector.
The vector sometimes may take big values and the inverse of it, may be close to
singularity.
4. Stability proof
which is non-negative since the matrices Q(x), R(x) and K(x) are positive-definite
and B(x)R(x) −1 B(x)𝑇 is also non-negative.
In summary, having derivative of Lyapunov function (23), the local exponen-
tially stability, and hence the local uniform asymptotic stability of the system (22)
follows [2, 7, 20]. The local stability result cannot be readily generalized to the
global stability, however, such generalization can be carried out through utilizing
the concept of region of attraction [14] for defined bounded set
¤
𝛀 = x ∈ R𝑛 , V(x) ¬ 𝑐, 𝑡 ∈ 𝑡 0 , 𝑡 𝑓 , (25)
where 𝑐 is a positive constant.
The nonlinear F16 aircraft model is applied to check the described finite-
time SDRE control for flight stabilization when the wind or other external forces
try to destabilize aircraft fligt-path. Governing equations that describe aircraft
aerodynamics are given by (5)-(6), but for the control purpose, state-dependent
parametrization SDC is necessary. When considering the flight dynamics for
566 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ
nonzero speed 𝑢 ≠ 0, parametrized F16 model (10) based on system (5) and (6)
with gravity compensation, can be described in SDC form
𝑢¤ (𝑔 sin 𝜃)/𝑢 0 0 0 −𝑤 𝑣
𝑢
(−𝑔 sin Ø cos 𝜃)/𝑢 −𝑢
𝑣¤ 0 0 𝑤 0 𝑣
𝑤¤ (−𝑔 cos Ø cos 𝜃)/𝑢 0 −𝑣
0 𝑢 0 𝑤
𝑝¤ =
0 0 𝑐 1 𝑝 + 𝑐 2𝑟
𝑝
0 0 0
𝑞¤
𝑐 4 𝑝 − 𝑐 3𝑟
𝑞
0 0 0 𝑐3 𝑝 0
𝑟¤
0 0 𝑐 5 𝑝 + 𝑐 6𝑟
𝑟
0 0 0
1/𝑚 0 0 0
0 0 0 0 𝐹𝑇
0 0 0 0 𝐿
+ , (26)
0 𝑐 𝑙𝑝 0 𝑐 𝑛𝑝 𝑀
𝑐 𝑚𝑞 𝑍𝑇 𝑃 0 𝑐 𝑚𝑞 0 𝑁
0 𝑐 𝑙𝑟 0 𝑐 𝑛𝑟
where control vector consists of thrust 𝐹𝑇 , and rolling, pitching, yawing moments
resulted from arrangements of ailerons, elevators and rudder [34].
As shown in Fig. 2, the thrust acts positively along the positive body 𝑥-axis.
Positive thrust cause an increase in acceleration along the body 𝑥-axis. For the
other control surfaces a positive deflection gives a decrease in the body rates.
A positive aileron deflection 𝛿𝑎 gives a decrease in the roll rates, this requires that
the right aileron deflect downward and the left aileron deflect upward. A positive
elevator deflection 𝛿 𝑒 results in a decrease in pitch rate, thus elevator is deflected
downwards. Positive deflection of the rudder 𝛿𝑟 decreases the yaw rate, and can
be described as a deflection to right. The maximum control values and units are
listed in Table 1.
Table 1: Minimum and maximum control values
presented in Table 1
𝑢¤ (𝑔 sin 𝜃)/𝑢 0 0 0 −𝑤 𝑣 𝑢
𝑣¤ (−𝑔 sin Ø cos 𝜃)/𝑢 −𝑢 𝑣
0 0 𝑤 0
𝑤¤ (−𝑔 cos Ø cos 𝜃)/𝑢 0 −𝑣
𝑤
0 𝑢 0
𝑝¤ =
0 0 0 0 𝑐 1 𝑝 + 𝑐 2𝑟 0 𝑝
𝑞¤
0 0 0 𝑐3 𝑝 0 𝑐 4 𝑝 − 𝑐 3𝑟 𝑞
𝑟¤
0 0 0 0 𝑐 5 𝑝 + 𝑐 6𝑟 0 𝑟
1/𝑚 0 0 0
0 0 0 0 𝐹𝑇
0 0 0 0 𝛿𝑒
+ , (27)
0 0 𝑐 𝑙 𝑝 𝑐 𝑙1 + 𝑐 𝑛𝑝 𝑐 𝑛1 𝑐 𝑙 𝑝 𝑐 𝑙2 + 𝑐 𝑛𝑝 𝑐 𝑛2 𝛿𝑎
𝑐 𝑚𝑞 𝑍𝑇 𝑃 𝑐 𝑚𝑞 𝑐 𝑚1
0 0 𝛿𝑟
0
0 𝑐 𝑙𝑟 𝑐 𝑙1 + 𝑐 𝑛𝑟 𝑐 𝑛1 𝑐 𝑙𝑟 𝑐 𝑙2 + 𝑐 𝑛𝑟 𝑐 𝑛2
important, because the aircraft should rapidly answer for pilot commands. The
path of flight must be sometimes rapidly stabilized when unexpected external
forces try to change the aircraft position and orientation during flying action.
Considering above, the problem consists of finding F16 aircraft state dynamics
and SDRE controls. In association with the aircraft dynamics (26), the quadratic
cost functional weighting matrices in (8) are chosen as
1 0.15 0 0 0 0
0.15 1 0 0 0 0
3 0 0 1 0 0 0
S = 5 · 10 , Q = 5 · 103 I6×6 .
0 0 0 1 0 0
0 0 0 0 1 0
0 0 0 0 0 1
and R = 2 · 10−4 I4×4 with initial speeds 𝑢 = 300 km/h, 𝑣 = −20 km/h, 𝑤 =
20 km/h and angle 𝜃 = 5◦ , in other hand, presenting it in vector form x0 =
𝑇
300 · 0.28 − 25 · 0.28 20 · 0.28 0 5 · 𝜋/180 0 , where 0.28 is a constant that
allows to recalculate the speed in km/h to m/s and 𝜋/180 allows to recalculate
degs to rads. The control should stabilize the aircraft in finite time 𝑡 𝑓 and uphold
linear speed 𝑢 on prescribed level, it means that the final state conditions should
𝑇
be as follows, for instance x 𝑓 = 500 · 0.28 0 0 0 0 0 . Considering the
final state, control law (13) takes the form
u = −R(x) −1 B(x)𝑇 K(x)(x − x 𝑓 ). (28)
Simulations are done to show the performance of the control designed in Section 3.
The aircraft state dynamics, controls and flight trajectory from initial state to the
final state are shown below.
Firstly, simulations are performed for the final time 𝑡 𝑓 = 3 s.
Next simulations are performed for the final time 𝑡 𝑓 = 5 s.
Figures 3–12 show closed-loop response of the flight controller and F16
aircraft system. Simulations are performed for two assumptions of the final control
time 𝑡 𝑓 : 3 and 5 seconds. Simulation time for both cases is two times longer than
control time 𝑡 𝑓 , because it is interesting haw the system works after 𝑡 𝑓 when
transversality condition holds.
When look at Figs. 3 and 8, it is worth to observe that for 𝑡 𝑓 = 3 s, the aircraft
is faster stabilized by feedback control and increase altitude 𝑧 only within 3.1
meters and distance 𝑥 approximately 202 meters. For the 𝑡 𝑓 = 5 s, the aircraft
is also stabilized well, but the aircraft change the altitude to 4.6 meters at the
distance 𝑥 equal to 585 meters. Hence the aircraft can be successively controlled
within assumed finite control times.
When consider controls, firstly thrust and aircraft moments shown at Figs. 4a–
4b and Figs. 9a–9b, further elevators, ailerons and rudder functions of deflection,
FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 569
Figure 4
570 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ
Figure 9
572 M. CHODNICKI, P. PIETRUSZEWSKI, M. WESOŁOWSKI, S. STĘPIEŃ
the maximum thrust is kept longer for shorted control time 𝑡 𝑓 = 3 s. Moments
and deflection functions are similar for both cased, except yawing moments and
rudder deflections. The thrust is kept longer for 𝑡 𝑓 = 3 s because the aircraft must
faster stabilize and straighten flight trajectory.
Analysis of the aircraft speeds, presented at Figs. 6 and 11, proof that speed
in the flight direction is increased to 500 km/h with simultaneously zeroing other
speeds for both cases. The situation is consistent to assumed starting and final
state conditions. Additionally, the solution of kinematics is presented at Figs. 7
and 12. There is shown how evaluates the aircraft space orientation.
6. Conclusions
The finite time control problem for flight stabilization in flight control system
in multirole F16 aircraft is formulated and solved. The nonlinear, state-dependent
parametrized model of the aircraft is proposed. The optimal control technique
with nonlinear feedback compensator for computation of the control input that
minimizes energy delivered to the aircraft system and energy lost, performing
stabilization task is analyzed. The effectiveness of presented technique is demon-
strated on a numerical example where optimal aircraft controls are found for
different final times. Presented results proof, that proposed SDRE technique can
be successively applied to the F16 flight control aircraft systems.
References
[2] H.T. Banks, B.M. Lewis, and H.T. Tran: Nonlinear feedback controllers
and compensators: a state-dependent Riccati equation approach. Com-
putational Optimization and Applications, 37(2), (2007), 177–218. DOI:
10.1007/s10589-007-9015-2.
[5] T. Cimen and S.P. Banks: Global optimal feedback control for general non-
linear systems with non-quadratic performance criteria. System and Control
Letters, 53(5), (2004), 327–346. DOI: 10.1016/j.sysconle.2004.05.008.
[6] T. Cimen and S.P. Banks: Nonlinear optimal tracking control with appli-
cation to super-tankers for autopilot design. Automatica, 40(11), (2004),
1845–1863. DOI: 10.1016/j.automatica.2004.05.015.
[7] T. Çimen: Systematic and effective design of nonlinear feedback controllers
via the state-dependent Riccati equation (SDRE) method. Annual Reviews
in Control, 34(1), (2010), 32–51. DOI: 10.1016/j.arcontrol.2010.03.001.
[8] J.R. Cloutier, C.N. D’Souza, and C.P. Mracek: Nonlinear regulation and
nonlinear H∞ control via the state-dependent Riccati equation technique:
Part 1, Theory; Part 2, Examples. In: Proceedings of the First International
Conference on Nonlinear Problems in Aviation and Aerospace, Daytona
Beach, FL, (1996), 117–141.
[9] R. Czyba and L. Stajer: Dynamic contraction method approach to digital
longitudinal aircraft flight controller design. Archives of Control Sciences,
29(1), (2019), 97–109. DOI: 10.24425/acs.2019.127525.
[10] E.B. Erdem and A.G. Alleyne: Globally stabilizing second-order nonlinear
systems by SDRE control. In: Proceedings of the American Control Con-
ference, San Diego, CA. USA, (1999). DOI: 10.1109/ACC.1999.786502.
[11] E.B. Erdem and A.G. Alleyne: Design of a class of nonlinear con-
trollers via state dependent Riccati equations. IEEE Transactions on Control
Systems Technology, 12(1), (2004), 133–137. DOI: 10.1109/TCST.2003.
819588.
[12] J. Geng, Y. Sheng, and X. Liu: Finite-time sliding mode attitude control
for a reentry vehicle with blended aerodynamic surfaces and a reaction
control system. Chinese Journal of Aeronautics, 27(4), (2014), 964–976.
DOI: 10.1016/j.cja.2014.03.013.
[13] K.D. Hammett, C.D. Hall, and D.B. Ridgely: Controllability issues in
nonlinear state-dependent Riccati equation control. Journal of Guidance,
Control and Dynamics, 21(54), (1998), 767–773. DOI: 10.2514/2.4304.
[14] A. Heydari and S.N. Balakrishnan: Closed-form solution to finite-
horizon suboptimal control of nonlinear systems. International Journal
of Robust Nonlinear Control, 25(15), (2015), 2687–2704. DOI: 10.1002/
rnc.3222.
[15] M.H. Korayem and S.R. Nekoo: Finite-time state-dependent Riccati equa-
tion for time-varying nonaffine systems: Rigid and flexible joint manip-
FINITE-TIME SDRE CONTROL OF F16 AIRCRAFT DYNAMICS 575